Talks.

Large language models have changed both sides of the social-engineering table. Attackers use them to industrialize pretexting; defenders can use them to run cyber deception at a scale and fidelity that was impossible a few years ago. This talk walks through the current state of AI-enhanced social engineering, the design principles behind believable LLM-driven decoys, and how blue teams can turn the same tools loose on adversaries — drawing on research from Project Ossian and behavior-science-driven awareness programs.